CN1779226A - Internal combustion engine ignition timing control device - Google Patents

Abstract

An ignition timing control system for internal combustion engines, which is capable of preventing occurrence of variation in engine speed and vibrations which can be caused due to variation in combustion state between the cylinders, to thereby improve drivability. The ECU 2 of the ignition timing control system 1 calculates a statistically processed value Pmi_ls#i according to the in-cylinder pressure Pcyl#i, and calculates an averaging target value Piav_cmd by weighted averaging of a minimum value Pmi_ls_minl and a second minimum value Pmi_ls_min2 within a predetermined range of the statistically processed value. The ECU 2 calculates a correction value DIGCMP#i by performing a limiting process for setting the optimization correction value DIGOP#i for control of the ignition timing to a limit value on the advanced side, on the averaging correction value DIGPIAV#i for causing the statistically processed values Pmi_ls#i to follow the averaging target value Piav_cmd, such that the statistically processed values Pmi_ls#i becomes maximum. The ignition timing IGLOG#i is calculated for each cylinder based on the correction value DIGCMP#i.

Description

Internal combustion engine ignition timing control device

technical field

The present invention relates to an ignition timing control device for an internal combustion engine that controls the ignition timing of an internal combustion engine based on cylinder internal pressure.

Background technique

Conventionally, a device described in JP-A-2003-262177 is known as an ignition timing control device for controlling the ignition timing of an internal combustion engine using a cylinder internal pressure sensor. In this ignition timing control device, the deviation between the maximum value of the cylinder internal pressure detected by the cylinder internal pressure sensor and the cylinder internal pressure at the top dead center of the compression stroke is calculated, and when the deviation is smaller than a predetermined threshold value, it is considered that The ignition timing is in the over-retarded state, and the ignition timing is controlled to a predetermined value on the advance side, and in other cases, the ignition timing is maintained at the value at that time.

Generally, in an internal combustion engine, there may be variations in the combustion state between cylinders (that is, there may be variations in the amount of heat generated by combustion), and in this case, torque variations will occur between cylinders, resulting in fluctuations in the rotational speed of the internal combustion engine, vibrations, etc. There is thus a risk of reduced running performance. Also, when an internal combustion engine is used as a power source of a vehicle, there is a risk of surge. Especially in the case where the internal combustion engine is respectively provided with a variable valve lift mechanism for continuously changing the lift of the intake valve for each cylinder, due to the difference in play and action characteristics of each variable valve lift mechanism, Such variations in the combustion state among the cylinders tend to occur.

On the other hand, according to the ignition timing control device of the above-mentioned Japanese Patent Laid-Open No. 2003-262177, since the ignition timing is controlled without considering the variation in the combustion state between the cylinders, a decrease in operability and occurrence of surge may occur. In the case of application to an internal combustion engine having a variable valve lift mechanism, there is an increased possibility that such a problem becomes conspicuous.

Contents of the invention

The present invention was made to solve the above-mentioned problems, and an object of the present invention is to provide an ignition timing control device for an internal combustion engine capable of improving operability by avoiding fluctuations in rotation speed and vibration due to variations in combustion states among cylinders.

In order to achieve the above object, the present invention provides an ignition timing control device for an internal combustion engine, which is characterized in that it has: a cylinder internal pressure detection device, which respectively detects the pressure in a plurality of cylinders as a plurality of cylinder internal pressures; A calculating device, which calculates a plurality of work quantity parameters representing work quantities caused by combustion in a plurality of cylinders according to the detected multiple cylinder pressures; a target value calculation device, which calculates the calculated multiple work quantities One of the quantity parameters is set as a reference value of a plurality of work quantity parameters, and based on the reference value, one target value that becomes a target of a plurality of work quantity parameters is calculated; and an ignition timing determining device determines a plurality of The ignition timing of the cylinders is such that multiple work parameters follow 1 target value.

According to the ignition timing control device of the internal combustion engine, the ignition timings of the plurality of cylinders are individually determined so that the plurality of work amount parameters follow one target value. In this case, since a plurality of work amount parameters represent work amounts caused by combustion in a plurality of cylinders, the ignition timing is controlled so that there is no variation in combustion state (i.e., variation in heat generation) among cylinders, thereby It is possible to avoid rotation speed fluctuations and vibrations due to variations in the combustion state between cylinders. Moreover, since one target value is not unrelated to multiple work parameters and is calculated based on one of the multiple work parameters, it is possible to avoid setting one target value as being related to all work parameters. The value of the parameter greatly deviates, so that the followability of multiple work parameters to a target value can be improved, and the occurrence of rotational speed fluctuations can be avoided. According to the above effects, the running performance can be improved. In particular, in the case where the ignition timing control device is applied to an internal combustion engine prone to combustion state deviation by providing a variable valve lift mechanism or the like for changing the lift of the intake valve, the above can be more effectively obtained. various advantages.

It is preferable to set the reference value to the minimum value among the plurality of work amount parameters.

According to the configuration of this preferred mode, since the reference value is set as the minimum value among the plurality of effort parameters, one target value to be the target of the plurality of effort parameters is calculated based on the minimum value, and a plurality of The ignition timing of the cylinders is such that multiple work parameters follow 1 target value. In this way, all the parameters of the amount of work including the parameter of the minimum value of the work amount can be reliably followed by one target value. As a result, for example, when the performance parameter of a certain cylinder shows a minimum value due to inter-individual variation or secular variation, etc., such a cylinder is not required to perform an impossible performance, thereby reliably avoiding multiple cylinders. The deviation of the combustion state. By the above means, the running performance can be further improved.

Preferably, the reference value is set to the minimum value of the work amount parameters within a range in which the maximum value of the work amount parameters is the upper limit and a value smaller than the maximum value by a predetermined value is the lower limit.

In general, in an internal combustion engine having a plurality of cylinders, when there is a cylinder whose combustion state is extremely deteriorated compared with other cylinders due to sludge or the like, when the ignition timing is controlled with the cylinder as a reference When the combustion state of other cylinders is consistent with this cylinder, in the cylinder with good combustion state, the ignition timing becomes an over-retarded state, which may cause high temperature of the exhaust gas and misfire, etc., causing the exhaust system and catalyst damaged. In contrast, according to the configuration of the preferred mode, since the reference value is set as the work amount within the range of taking the maximum value among the multiple work amount parameters as the upper limit and taking a value smaller than the maximum value by a predetermined value as the lower limit According to the minimum value of the parameters, one target value of multiple work parameters is calculated, so even if the difference between the maximum value and the minimum value of multiple work parameters is greater than the specified value, that is, there is combustion In the case of a cylinder whose state is extremely worse than other cylinders, the ignition timing can be controlled so that there is no combustion state deviation based on cylinders other than the cylinder whose deviation from the maximum value of the work parameter is equal to or less than a predetermined value. As a result, the temperature rise of the exhaust gas, the occurrence of misfires, etc. can be avoided, and damage to the exhaust system and the catalyst can be avoided.

Preferably, the work parameter calculation device calculates the unprocessed values of the multiple work parameters based on the multiple cylinder internal pressures, and calculates the multiple work parameters by performing predetermined sequential statistical processing on the unprocessed values.

In general, the internal cylinder pressure fluctuates randomly, so if the detected internal cylinder pressure is used as it is, the work amount parameter calculated based thereon will also become a value with noise due to its influence. As a result, when the ignition timing is controlled using such a work amount parameter, the control accuracy of the ignition timing may decrease accordingly due to the influence of noise. In contrast, according to the configuration of the preferred mode, since the plurality of work parameters are calculated by performing predetermined ordinal statistical processing on the unprocessed values of the plurality of work parameters calculated from the plurality of cylinder internal pressures, Therefore, a plurality of work amount parameters are calculated as values that are less affected by fluctuations in cylinder internal pressure, that is, values that suppress the influence of noise due to fluctuations, and the ignition timing can be controlled using such work amount parameters. As a result, the control accuracy of the ignition timing can be improved accordingly by suppressing the influence of noise caused by the variation in the cylinder internal pressure.

Preferably, the ignition timing determining means has: a first control value calculating means for calculating the first control value for controlling the ignition timing of each cylinder in each of the plurality of cylinders so that the work amount parameter of each cylinder becomes the maximum ; the second control value calculation device, which calculates the ignition timing for controlling each cylinder in each cylinder, so that the work amount parameter of each cylinder follows the second control value of a target value; the final control value calculation device, which passes The second control value is subjected to limit processing with the first control value being the limit value on the advance angle side to calculate a final control value; and an ignition timing calculation device that calculates the ignition timing of each cylinder based on the calculated final control value .

According to the configuration of this preferred mode, in each cylinder, the first control value is calculated as the value used to control the ignition timing so that the work amount parameter of the cylinder becomes the maximum value, and the second control value is calculated as the value used to control the ignition timing. The work quantity parameter of the cylinder is calculated following a target value, and the final control value is calculated by performing limit processing on the second control value with the first control value as the limit value on the advance angle side. According to the final control value to calculate the ignition timing for each cylinder. In this way, since the final control value is calculated by performing limit processing on the second control value with the first control value being the limit value on the advance angle side, the ignition timing for maximizing the work amount parameter is executed simultaneously with each cylinder. The timing control is different from the ignition timing control used to make the work parameter follow one target value, which can avoid the interference of the two ignition timing controls, and can smoothly switch the final control value from the second control value to the second control value. 1 Switching at control value. In this way, the occurrence of torque steps can be avoided, and the running performance can be improved. Furthermore, for the same reason, in any cylinder, even when the second control value is set to a value closer to the advance angle side than the first control value, the final control value of the cylinder can be limited to The first control value is used to calculate the ignition timing based on the first control value, so that the work parameter of the cylinder can be maximized. That is, the highest combustion efficiency can be ensured. In this way, the running performance can be further improved.

More preferably, the target value calculating means calculates one target value by weighted average calculation of a reference value and a value larger than the reference value among the plurality of work amount parameters.

According to the configuration of this preferred mode, since one target value is calculated by weighted average calculation of a reference value and a value larger than the reference value in a plurality of work parameters, a value larger than the reference value is calculated as a target value, Therefore, in the cylinder whose work quantity parameter is the reference value, the probability that the final control value is set as the first control value and the ignition timing is controlled so that the work quantity parameter becomes the maximum increases. As a result, in the cylinder of the reference value, since the ignition timing is controlled so that the work amount parameter becomes the maximum, the combustion efficiency of the internal combustion engine as a whole can be improved. In particular, as in the above-mentioned second preferred mode, the reference value is set to be one of the work parameters in the range in which the maximum value among the multiple work parameters is set as the upper limit and the value smaller than the maximum value by a predetermined value is the lower limit. In the case of the minimum value, in addition to the cylinder with the reference value, even in the cylinder whose work parameter is smaller than the reference value, the ignition timing is controlled so that the work parameter is the largest, so that the combustion efficiency of the internal combustion engine as a whole can be improved more effectively .

More preferably, the ignition timing determining means has: a first control value calculating means which calculates a first control value for controlling the ignition timing of each of the plurality of cylinders so that the work amount parameter of each cylinder is maximized. value; the 2nd control value calculation means, which calculates the 2nd control value for controlling the ignition timing of each cylinder in each cylinder so that the work amount parameter of each cylinder follows a target value; and the ignition timing calculation means, It calculates the ignition timing of cylinders whose work parameters are less than or equal to the reference value based on the first control value, and calculates the ignition timing of cylinders whose work parameters are greater than the reference value based on the second control value.

According to the configuration of this preferred mode, in each cylinder, the first control value is calculated as the value used to control the ignition timing so that the work amount parameter of the cylinder becomes the maximum value, and the second control value is calculated as the value used to control the ignition timing. The work parameter of the cylinder is calculated following a target value, the ignition timing of the cylinder whose work parameter is less than or equal to the reference value is calculated according to the first control value, and the cylinder whose work parameter is greater than the reference value is calculated according to the second control value The ignition timing of the cylinder. By calculating the ignition timing in this way, in the cylinder whose work parameter is less than or equal to the reference value, it is controlled so that the work parameter of the cylinder is the largest, and in the cylinder whose work parameter is greater than the reference value, it is controlled so that the cylinder’s The amount of work parameter follows a target value. In this way, even in the case where there is a cylinder whose combustion state is extremely deteriorated compared with the cylinder of the reference value, in the cylinder, the maximum value of the work amount parameter can be ensured, and in the case of the cylinder whose work amount parameter is the reference value and the work amount In the cylinders whose parameters are larger than the reference value, it is possible to ensure the work amount parameter equivalent to the maximum value of the work amount parameter in the cylinder of the reference value while eliminating the variation in the combustion state. As a result, the combustion efficiency of the entire internal combustion engine can be more effectively improved.

More preferably, the second control value calculating means calculates the second control value using a control algorithm including a two-degree-of-freedom control algorithm combining a predetermined target value filter algorithm and a predetermined follow-up control algorithm.

Since the second control value is a value for controlling the ignition timing of each cylinder so that each work parameter follows a target value, when the operating state of the internal combustion engine changes, the cylinder that becomes the reference value of the work parameter is switched. When , there is a possibility that the second control value also changes suddenly due to a sudden change of one target value. In contrast, according to the configuration of this preferred mode, since the second control value is calculated using a control algorithm including a 2-degree-of-freedom control algorithm that combines a predetermined target value filtering algorithm and a predetermined follow-up control algorithm, by appropriately Set the filtering characteristics of the target value filtering algorithm accurately, even if one target value changes suddenly, it can avoid the sudden change of the second control value. As a result, it is possible to avoid the occurrence of a torque step caused by sudden changes in the second control value, and to further improve the drivability.

The above and other objects, features, and advantages of the present invention will become more apparent from the following detailed description based on the accompanying drawings, in which the same symbols denote the same components in the drawings.

Description of drawings

FIG. 1 is a schematic diagram showing a schematic configuration of an internal combustion engine to which an ignition timing control device according to an embodiment of the present invention is applied.

FIG. 2 is a schematic diagram showing a schematic configuration of an ignition timing control device.

FIG. 3 is a block diagram showing a schematic configuration of an ignition timing control device.

4 is a timing chart showing changes in estimated value Pmi_act#i and statistically processed value Pmi_ls#i accompanying changes in cylinder internal pressure Pcyl#i during a combustion cycle.

FIG. 5 is a diagram showing an example of a table used for calculation of the identification signal value DIGID.

FIG. 6 is a diagram showing an example of a map used for calculation of the basic ignition timing IGBASE.

FIG. 7 is a flowchart showing ignition timing control processing.

FIG. 8 is a diagram showing an example of a table used for setting the cylinder number value #i.

FIG. 9 is a flowchart showing Pmi optimization and smoothing control processing.

FIG. 10 is a flowchart showing intake air amount control processing.

FIG. 11 is a diagram showing an example of a table used for calculation of the start-time value Gcyl_cma_crk of the target intake air amount.

FIG. 12 is a diagram showing an example of a map used for calculation of the target valve lift Liftin_cmd.

FIG. 13 is a diagram showing an example of a map used for calculating the catalyst warm-up value Gcyl_cmd_ast of the target intake air amount.

FIG. 14 is a diagram showing an example of a map used for calculation of the normal-time value Gcyl_cmd_drv of the target intake air amount.

15 is a timing chart showing an example of a simulation result of ignition timing control by the ignition timing control device.

FIG. 16 is a timing chart showing an example of a simulation result of ignition timing control by the ignition timing control device of the modified example.

17 is a timing chart showing an example of a simulation result of ignition timing control by the ignition timing control device of the comparative example.

Detailed ways

Hereinafter, an ignition timing control device for an internal combustion engine according to an embodiment of the present invention will be described with reference to the drawings. As shown in FIG. 2, the ignition timing control device 1 has an ECU (Electronic Control Unit, electronic control unit) 2. As will be described later, the ECU 2 executes the ignition timing according to the operating state of the internal combustion engine (hereinafter referred to as "engine") 3. Time control and other control processing.

As shown in FIG. 1 , the engine 3 is an in-line 4-cylinder gasoline engine mounted on a vehicle not shown, and has 1st to 4th cylinders #1 to #4 (plurality of cylinders). The intake pipe 4 of the engine 3 is connected to the four cylinders #1 to #4 through four branch portions 4b of an intake manifold 4a.

In each branch portion 4b, a fuel injection valve 5 is attached on the upstream side of an intake port (not shown) of each cylinder. When the engine 3 is running, the valve opening time and the valve opening timing of each fuel injection valve 5 are controlled by a drive signal from the ECU 2. In this way, fuel injection control is performed.

In addition, a spark plug 6 is attached to each cylinder on a cylinder head of the engine 3 . The spark plug 6 is connected to the ECU 2 through an ignition coil not shown in the figure, and the spark plug 6 is discharged by applying a drive signal (voltage signal) from the ECU 2 at a timing corresponding to the ignition timing IGLOG described later, so that the combustion chamber combustion of the mixture.

Furthermore, a variable valve lift mechanism 10 is provided in the engine 3 . The variable valve lift mechanism 10 changes the intake air amount steplessly by changing the lift of an intake valve (hereinafter referred to as "valve lift") Liftin, not shown, steplessly within a predetermined range. The detailed description thereof is omitted here, and the structure is the same as that of the patent application No. 2004-249213 already filed by the present applicant (assignee). The variable valve lift mechanism 10 is electrically connected to the ECU 2, and is driven by a lift control input Uliftin from the ECU 2 to change the valve lift Liftin. In addition, in the present embodiment, it is assumed that the valve lift Liftin represents the maximum lift of the intake valve 4 .

On the other hand, as shown in FIG. 2 , the ECU 2 communicates with a crank angle sensor 20, a water temperature sensor 21, a cylinder internal pressure sensor 22, a valve lift sensor 23, an accelerator opening sensor 24, and an ignition switch (hereinafter referred to as "IG/ SW")25 connection.

The crank angle sensor 20 outputs the CRK signal and the TDC signal, both of which are pulse signals, to the ECU 2 as the crankshaft (not shown) rotates. The CRK signal outputs one pulse for each predetermined crank angle (1° in this embodiment), and the ECU 2 calculates the rotational speed (hereinafter referred to as "engine rotational speed") NE of the engine 3 based on the CRK signal. Furthermore, the TDC signal is a signal indicating that the piston of each cylinder is located at a predetermined crank angle position slightly ahead of the TDC position of the intake stroke, and one pulse is output for each predetermined crank angle.

Further, the water temperature sensor 21 is composed of a thermistor or the like mounted on a cylinder block (not shown) of the engine 3, and outputs a detection signal of the engine water temperature TW, which is the temperature of cooling water circulating in the cylinder body, to the ECU. 2.

Further, the cylinder internal pressure sensor 22 (cylinder internal pressure detecting device) is a piezoelectric element type sensor integrated with the spark plug 6 and is provided for each cylinder (only one is shown in the figure). The cylinder internal pressure sensor 22 is deflected according to changes in the internal cylinder pressure Pcyl#i, which is the pressure in each cylinder, and outputs a detection signal indicating the cylinder internal pressure Pcyl#i to the ECU 2. In addition, #i (=#1 to #4) in the "in-cylinder pressure Pcyl#i" is a cylinder number value indicating the cylinder number. For example, the cylinder internal pressure Pcyl#1 corresponds to the first cylinder #1 pressure. This point is the same for various parameters described later.

In addition, the valve lift sensor 23 outputs a detection signal indicating the valve lift Liftin to the ECU 2, and the accelerator opening sensor 24 outputs a depressing amount (hereinafter referred to as "accelerator opening") indicating the amount of depression of the accelerator pedal (not shown) of the vehicle. ) The detection signal of the AP is output to the ECU 2. Also, the IG/SW 25 is turned on/off by operation of an ignition key (not shown), and outputs a signal indicating its on/off state to the ECU 2 .

On the other hand, the ECU 2 is composed of a microcomputer including a CPU, RAM, ROM, and I/O interfaces (all not shown in the figure). On/off signals and the like are used to determine the operating state of the engine 3, and various control processes including ignition timing control are executed as will be described later.

In addition, in the present embodiment, the ECU 2 corresponds to the work parameter calculation means, the target value calculation means, the ignition timing determination means, the first control value calculation means, the second control value calculation means, the final control value calculation means and the ignition timing determination means. Timing calculation device.

Next, the ignition timing control device 1 according to the present embodiment will be described with reference to FIG. 3 . As shown in FIG. 3 , the ignition timing control device 1 includes: an estimated value calculation unit 30 , a statistically processed value calculation unit 31 , a smoothing target value calculation unit 32 , a smoothing controller 33 , a signal value calculation unit 34 for identification, an MBT The estimation unit 35, the optimization controller 36, the correction value calculation unit 37, the basic ignition timing calculation unit 38, and the adder 39, specifically, all of them are constituted by the ECU 2. In this ignition timing control device 1, the ignition timing IGLOG#i is calculated as described below.

First, in the estimated value calculation unit 30, an estimated value (hereinafter referred to as "estimated value") Pmi_act#i of indicated mean effective pressure (Indicated Mean Effective Pressure, IMEP) is calculated using the spectral decomposition method. Since this spectral decomposition method is a method proposed by the applicant (assignee) in Patent Application No. 2004-232633, its detailed description is omitted here. Utilizing this spectral decomposition method, according to the The estimated value Pmi_act#i is calculated using the following equation (1) for the cylinder internal pressure Pcyl#i.

$\mathrm{<span\; class="notranslate">\; Pmi</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; act</span>}<span\; class="notranslate">\; \#</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{{\mathrm{<span\; class="notranslate">\; V</span>}}_{\mathrm{<span\; class="notranslate">\; the\; s</span>}}}\underset{\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; N</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\mathrm{<span\; class="notranslate">\; PF</span>}<span\; class="notranslate">\; \#</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \{</span>\mathrm{<span\; class="notranslate">\; VF</span>}<span\; class="notranslate">\; \#</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; VF</span>}<span\; class="notranslate">\; \#</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; no</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \}</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>$

In the formula, Vs is the stroke volume of the engine 3 , and PF#i and VF#i represent the values obtained by subjecting the internal pressure Pcyl#i and the volume data V#i of each cylinder to a predetermined low-pass filter process, respectively. Also, each data marked with (n) represents discrete data that is sampled (or calculated) in synchronization with the generation of the CRK signal, and the symbol n represents the order of the sampling period of each discrete data. For example, symbol n represents a value sampled at the current control timing, and symbol n-1 represents a value sampled at the previous control timing. Furthermore, symbol N represents the number of data sampled in one combustion cycle (value 720 in this embodiment). And, discrete data with a symbol (k) means data that is calculated or sampled in synchronization with the occurrence of the TDC signal. The above-mentioned points are also the same for each discrete data described later. In addition, in the following description, the symbol (k) etc. of each discrete data are omitted suitably.

According to the spectral decomposition method described above, estimated value Pmi_act#i is calculated using values PF#i and VF#i obtained by performing predetermined low-pass filter processing on in-cylinder pressure Pcyl#i and volume data V#i. The relative phase deviation between the cylinder internal pressure Pcyl#i and the volume data V#i is avoided, so that the calculation accuracy of the estimated value Pmi_act#i can be improved. In addition, in the present embodiment, the estimated value calculation unit 30 corresponds to the work amount parameter calculation means and the second control value calculation means, and the estimated value Pmi_act#i corresponds to the unprocessed value.

Then, the statistically processed value calculation unit 31 calculates the statistically processed value Pmi_ls#i indicating the mean effective pressure by performing statistical processing described below on the estimated value Pmi_act#i. Specifically, the statistical processing value Pmi_ls#i is calculated using the sequential least square method algorithm of the fixed gain method represented by the following equations (2) and (3).

$\mathrm{<span\; class="notranslate">\; Pmi</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; ls</span>}<span\; class="notranslate">\; \#</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; Pmi</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; ls</span>}<span\; class="notranslate">\; \#</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; +</span>\frac{\mathrm{<span\; class="notranslate">\; <figure-callout\; id="1"\; label="Pav"\; filenames="A20051012435200351.png,A20051012435200391.png"\; state="\{\{state\}\}">Pav</figure-callout></span>}}{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; Pav</span>}}\mathrm{<span\; class="notranslate">\; E.</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; id</span>}<span\; class="notranslate">\; \#</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; )</span>$

E_id#i(k)＝Pmi_act#i(k)-Pmi_ls#i(k-1) .....(3)

In the above formula (2), Pav is a predetermined filter gain, and E_id#i is a deviation calculated using formula (3).

Comparing the statistically processed value Pmi_ls#i calculated as described above with the estimated value Pmi_act#i shows that, as shown in FIG. The statistical processing described above suppresses the influence even when the in-cylinder pressure Pcyl#i fluctuates during the combustion cycle.

In addition, in the present embodiment, the statistically processed value calculating unit 31 corresponds to the work amount parameter calculating means and the second control value calculating means, and the statistically processed value Pmi_ls#i corresponds to the work amount parameter.

And in the smoothing target value calculation part 32, the smoothing target value Piav_cmd is calculated by the weighted average calculation shown by following Formula (4). This smoothing target value Piav_cmd is used to smooth the indicated average effective pressures of the four cylinders #1 to #4, that is, the amount of work.

Piav_cmd(k)＝w1·Pmi_ls_min1(k)+w2·Pmi_ls_min2(k) .....(4)

In the formula, w1 and w2 are weighting coefficients, which are set to satisfy predetermined values satisfying w1+w2=1, 0≤w1, and 0≤w2. Pmi_ls_min1 and Pmi_ls_min2 are the first and second reference values. Specifically, when the maximum processing value among the four statistical processing values Pmi_ls#i is taken as the maximum value Pmi_ls_max and Eps_Pmi is a positive predetermined value, the first reference The value Pmi_ls_min1 is set to the minimum value among statistically processed values satisfying (Pmi_ls_max−Eps_Pmi)≦Pmi_ls_#i. Furthermore, the second reference value Pmi_ls_min2 is set to the second smallest value among statistically processed values satisfying (Pmi_ls_max−Eps_Pmi)≦Pmi_ls#i.

In addition, in the present embodiment, the smoothing target value calculation unit 32 corresponds to a target value calculation device, and the smoothing target value Piav_cmd corresponds to one target value.

Furthermore, in the smoothing controller 33, the smoothing correction value DIGPIAV#i is calculated as described below. First, when the later-described limit value Lim_piav(Usl_piav#i) satisfies Lim_piav(Usl_piav#i(k-1))=Usl_piav#i(k-1), use the following equations (5) to (9) The 2-degree-of-freedom control algorithm calculates the smoothing control input value Usl_piav#i. This two-degree-of-freedom control algorithm is an algorithm obtained by combining the target value filtering algorithm shown in equation (9) and the sliding mode control algorithm (sliding mode control algorithm) shown in equations (5) to (8).

Usl_piav#i(k)=Krcn_piav·σ_piav#i(k)

+Kadp_piav#i(k) sum_σ_piav#i(k) .....(5)

$\mathrm{<span\; class="notranslate">\; sum</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; \&sigma;</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; piav</span>}<span\; class="notranslate">\; \#</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; k</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\mathrm{<span\; class="notranslate">\; \&sigma;</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; piav</span>}<span\; class="notranslate">\; \#</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 6</span>}<span\; class="notranslate">\; )</span>$

σ_piav#i(k)＝EPIAV#i(k)+S_piav·EPIAV#i(k-1)                                                                                          ...

EPIAV#i(k)＝Pmi_ls#i(k)-Piav_cmd_f(k) .....(8)

Piav_cmd_f(k)＝-R_piav·Piav_cmd_f(k-1)+(1+R_piav)Piav_cmd(k) .....(9)

In the above formula (5), Krch_piav#i and Kadp_piav#i are predetermined arrival law gain (arrival law ゲイソ) and adaptive law gain (adaptive law ゲイソ), respectively. Also, σ_piav #i is the switching function calculated using Expression (7), and sum_σ_piav #i is the integral value of the switching function calculated using Expression (6). In Equation (7), S_piav is a switching function setting parameter set to satisfy -1<S_piav<0, and EPIAV#i is a deviation defined by Equation (8). In this equation (8), Piav_cmd_f is a filter value of the smoothing target value Piav_cmd, and is calculated using the primary delay filter algorithm shown in equation (9). In this formula (9), R_piav is a target value response designation parameter set to satisfy -1<R_piav<0.

Then, from the smoothing control input value Usl_piav#i calculated as described above, the smoothing correction value DIGPIAV#i is calculated using the following equation (10).

DIGPIAV#i(k)＝Lim_piav(Usl_piav#i(k)) .....(10)

In the formula, Lim_piav(Usl_piav#i(k)) represents the limit value after limit processing is performed on the smoothing control input value Usl_piav#i(k). Specifically, Usl_piav#i(k) is limited to It is calculated from a value within the predetermined range defined by the predetermined lower limit value DIGPIAV_L and the predetermined upper limit value DIGPIAV_H. That is, when Usl_piav#i(k)<DIGPIAV_L, Lim_piav(Usl_piav#i(k))=DIGPIAV_L, when DIGPIAV_L≤Usl_piav#i(k)≤DIGPIAV_H, Lim_piav(Usl_piav#i(k))=Usl_piav# i(k), when Usl_piav#i(k)>DIGPIAV_H, Lim_piav(Usl_piav#i(k))=DIGPIAV_H.

In the above, the case where Lim_piav(Usl_piav#i(k-1))=Usl_piav#i(k-1) is established is the case where DIGPIAV_L≤Usl_piav#i(k-1)≤DIGPIAV_H is established, and in this case, the above formula is used (5) to (10) Calculate the smoothing correction value DIGPIAV#i.

On the other hand, when Lim_piav(Usl_piav#i(k-1))≠Usl_piav#i(k-1), that is, in the calculation process of the previous smoothing correction value DIGPIAV#i, Lim_piav(Usl_piav#i( When k-1))=DIGPIAV_L or Lim_piav(Usl_piav#i(k-1))=DIGPIAV_H, the following equation (11) is used instead of the above-mentioned equation in the calculation process of the smoothing control input value Usl_piav#i this time (6). That is, the smoothing control input value Usl_piav #i is calculated using equations (5), (7) to (11), and the update of the integral value sum_σ_piav #i of the switching function is stopped.

sum_σ_piav#i(k)＝sum_σ_piav#i(k-1)                                                                                                                                                                                                                                                ... (11)

Thus, when Lim_piav(Usl_piav#i)≠Usl_piav#i, updating of the integral value sum_σ_piav#i of the switching function is aborted for the following reason. That is, when Usl_piav#i is a value out of the above-mentioned predetermined range and is limited to DIGPIAV_L or DIGPIAV_H, if the update of the integral value sum_σ_piav#i of the switching function is continued, the value increases. As a result, when the smoothing control input value Usl_piav#i returns to the above-mentioned predetermined range, it takes a long time until the increased integral value sum_σ_piav#i of the switching function returns to an appropriate value, during which time the control accuracy decreases, This therefore needs to be avoided.

In addition, in the present embodiment, the smoothing controller 33 corresponds to the ignition timing determining means and the second control value calculating means, and the smoothing correction value DIGPIAV#i corresponds to the second control value.

On the other hand, in the identification signal value calculation unit 34 (first control value calculation means), as described below, the identification is calculated using the method proposed by the present applicant (assignee) in Patent Application No. Use the signal value DIGID. That is, as will be described later, the identification signal value DIGID is calculated by searching the table shown in FIG. 5 based on the counter value Cdigid. The identification signal value DIGID is set as a signal value satisfying the self-excitation condition for appropriately identifying the MBT deviation EIGOP in the identification calculation of the MBT estimation unit 35 described later, more specifically, it is set to include Fluctuating signal values or random signal values of more than 3 sine waves.

In addition, the MBT estimation unit 35 (first control value calculation means) calculates the MBT deviation EIGOP#i using the method proposed by the present applicant (assignee) in Patent Application No. 2003-385741 as described below. The MBT deviation EIGOP#i represents the deviation between the estimated MBT, which is an estimated value of the MBT (Minimum advance for Best Torque: the minimum ignition advance angle for best torque) in each cylinder, and the current set ignition timing IGLOG, that is, The advance amount from the currently set ignition timing IGLOG to the estimated MBT is specifically calculated using the following equation (12).

$\mathrm{<span\; class="notranslate">\; EIGOP</span>}<span\; class="notranslate">\; \#</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; Bigop</span>}<span\; class="notranslate">\; \#</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; \&Center\; Dot;</span>\mathrm{<span\; class="notranslate">\; Aigop</span>}<span\; class="notranslate">\; \#</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 12</span>}<span\; class="notranslate">\; )</span>$

In this formula (12), Aigop#i and Bigop#i are coefficients of an estimated MBT characteristic curve (quadratic curve) described later, and a coefficient vector having Aigop#i and Bigop#i as elements is calculated as follows θ#i.

First, the coefficient vector θ#i is defined by the following equation (13).

θ#i ^T (k)＝[Aigop#i(k), Bigop#i(k), Cigop#i(k)] .....(13)

First, use the fixed gain algorithm with the δ correction method shown in the following formulas (14)~(23), sequentially identify the unprocessed coefficient vector θ#i', and implement the formulas shown in formulas (24) and (25) Limit processing is performed to calculate the coefficient vector θ#i.

θ#i ’(k) = θ_base+dθ#i (k) ..... (14)

dθ#i(k)＝δ·dθ#i(k-1)+KP#i(k)·Eigop#i(k) .....(15)

$\mathrm{<span\; class="notranslate">\; KP</span>}<span\; class="notranslate">\; \#</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; P</span>}<span\; class="notranslate">\; \&Center\; Dot;</span>\mathrm{<span\; class="notranslate">\; \&zeta;</span>}<span\; class="notranslate">\; \#</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span>}{\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; +</span>\mathrm{<span\; class="notranslate">\; \&zeta;</span>}{<span\; class="notranslate">\; \#</span>\mathrm{<span\; class="notranslate">\; i</span>}}^{\mathrm{<span\; class="notranslate">\; T</span>}}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; \&Center\; Dot;</span>\mathrm{<span\; class="notranslate">\; P</span>}<span\; class="notranslate">\; \&Center\; Dot;</span>\mathrm{<span\; class="notranslate">\; \&zeta;</span>}<span\; class="notranslate">\; \#</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span>}<span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 16</span>}<span\; class="notranslate">\; )</span>$

Eigop#i(k)＝Pmi_act#i(k)-Pmi_hat#i(k) .....(17)

Pmi_hat#i(k)=θ#i' ^T (k-1)·ζ#i(k)

=Aigop#i'(k-1)·DIGID(k) ²

  +Bigop#i(k-1)·DIGID(k)+Cigop#i(k-1)                                                                                                                                                                                                              …

θ# ^i'T (k)＝[Aigop#i'(k), Bigop#i(k), Cigop#i(k)] .....(19)

dθ#i ^T (k)=[Aigop#i'(k)-Aigop_base, dBigop#i'(k), dCigop#i'(k)]

.....(20)

θ_base ^T (k) = [Aigop_base, 0, Cigop_base] .....(21)

ζ ^T (k) = [DIGID(k) ² , DIGID(k), 1] .....(22)

$\mathrm{<span\; class="notranslate">\; \&delta;</span>}<span\; class="notranslate">\; =</span>\left[\begin{array}{ccc}\mathrm{<span\; class="notranslate">\; 1</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}\\ \mathrm{<span\; class="notranslate">\; 0</span>}& {\mathrm{<span\; class="notranslate">\; \&delta;</span>}}^{<span\; class="notranslate">\; \&prime;</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}\\ \mathrm{<span\; class="notranslate">\; 0</span>}& \mathrm{<span\; class="notranslate">\; 0</span>}& {\mathrm{<span\; class="notranslate">\; \&delta;</span>}}^{<span\; class="notranslate">\; \&prime;</span>}\end{array}\right]<span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; twenty\; three</span>}<span\; class="notranslate">\; )</span>$

Θ#i (k) = lim_a (θ#i ’(k) ..... (24)

Lim_a(θ#i'(k)) ^T ＝[Lim_a(Aigop#i'(k)), Bigop#i(k), Cigop#i(k)]

.....(25)

In the above formula (14), the transposition matrix of the unprocessed coefficient vector θ#i' is defined according to formula (19), where Aigop#i' in the formula (19) is the coefficient Aigop before the limit processing described later Identification value of #i. Also, θ_base in Equation (14) is a reference value vector in which the transpose matrix is defined in Equation (21), and dθ#i is a correction term vector in which the transpose matrix is defined in Equation (20). This correction term vector dθ#i is calculated using Equation (15), and δ in Equation (15) is a forgetting vector defined by Equation (23). δ' in the equation (23) is a forgetting coefficient, and is set so as to satisfy 0<δ'<1.

Also, KP#i in Equation (15) is a vector of gain coefficients calculated using Equation (16). In this equation (16), P is a predetermined fixed gain (constant value), and ζ is a vector of identification signal values in which a transpose matrix is defined by equation (22). In addition, Eigop in Equation (15) is a deviation calculated using Equation (17), and Pmi_hat in Equation (17) is an identification value indicating the mean effective pressure, and is calculated by Equation (18). This equation (18) represents an estimated MBT curve (quadratic curve).

Then, using the unprocessed coefficient vector θ#i' calculated as described above, the coefficient vector θ#i is calculated using the above equation (24). Lim_a(θ#i') of the formula (24) is the limit processing value of the unprocessed coefficient vector θ#i', and its transpose matrix is defined according to the formula (25). Lim_a(Aigop#i') in the formula (25) is the limit processing value of Aigop#i', which is processed to a value that must satisfy Lim_a(Aigop#i')<0, and for formula (24), ( 25) is compared with formula (19), it can be seen that Lim_a(Aigop#i')=Aigop#i. That is, the coefficient Aigop#i is recognized as necessarily having a negative value. This is because, when the MBT deviation EIGOP#i is close to a value of 0, Aigop#i may be misrecognized as a positive value that makes the estimated MBT characteristic curve of Equation (18) a downwardly convex curve, so it is necessary to avoid This misidentification.

In addition, the optimization controller 36 calculates the optimization correction value DIGOP#i using the MBT deviation EIGOP#i calculated by the MBT estimation unit 35 as described below. First, when the limit value Lim_igop(Usl_igop#i) described later satisfies Lim_igop(Usl_igop#i(k-1))=Usl_igop#i(k-1), use the following equations (26) to (28) The sliding mode control algorithm calculates the control input value Usl_igop#i for optimization.

Usl_igop(k)＝Krch_igop·σ_igop#i(k)+Kadp_igop·sum_σ_igop#i(k) .....(26)

$\mathrm{<span\; class="notranslate">\; sum</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; \&sigma;</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; igop</span>}<span\; class="notranslate">\; \#</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\underset{\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; k</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\mathrm{<span\; class="notranslate">\; \&sigma;</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; igop</span>}<span\; class="notranslate">\; \#</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 27</span>}<span\; class="notranslate">\; )</span>$

σ_igop#i(k)＝EIGOP#i(k)+S_igop·EIGOP#i(k-1)                                                                                                                                                                                                                                                     ...(28)

In the above formula (26), Krcn_igop#i and Kadp_igop#i are the prescribed arrival law gain and adaptive law gain, respectively. Also, σ_igop #i is the switching function calculated using Equation (28), and susum_σ_igop #i is the integral value of the switching function calculated using Equation (27). In Equation (28), S_igop is a switching function setting parameter set to satisfy -1<S_igop<0.

Then, based on the optimization control input value Usl_igop#i calculated as described above, the optimization correction value DIGOP#i is calculated using the following equation (29). In addition, this optimal correction value DIGOP#i is calculated as a value obtained by correcting the basic ignition timing IGBASE described later to the advance angle side. This is because, in general, the base ignition timing IGBASE is set to a value on the retard side of MBT.

DIGOP#i(k)＝Lim_igop(Usl_igop(k)) .....(29)

In the formula, Lim_igop(Usl_igop#i(k)) represents the limit value after limiting the control input value Usl_igop#i(k) for optimization, specifically, Usl_igop#i(k) is limited as The value within the predetermined range defined by the predetermined lower limit value DIGOP_L and the predetermined upper limit value DIGOP_H is calculated.

That is, when Usl_igop#i(k)<DIGOP_L, Lim_igop(Usl_igop#i(k))=DIGOP_L, when DIGOP_L≤Usl_igop#i(k)≤DIGOP_H, Lim_igop(Usl_igop#i(k))=Usl_igop# i(k), when Usl_igop#i(k)>DIGOP_H, Lim_igop(Usl_igop#i(k))=DIGOP_H. In addition, the lower limit value DIGOP_L is set to a value larger than the above-mentioned lower limit value DIGPIAV_L, and the upper limit value DIGOP_H is set to a value smaller than the above-mentioned upper limit value DIGPIAV_H.

As above, the case where Lim_igop(Usl_igop#i(k-1))=Usl_igop#i(k-1) holds true is the case where DIGOP_L≤Usl_igop#i(k-1)≤DIGOP_H holds true, and in this case, the above formula is used (26) to (29) Calculate the optimization correction value DIGOP#i.

On the other hand, when Lim_igop(Usl_igop#i(k-1))≠Usl_igop#i(k-1), or when DIGOP#i(k-1)≥DIGPIAV#i(k-1), in this In the calculation process of the next optimization control input value Usl_igop#i(k), the following formula (30) is used instead of the above formula (27). That is, the optimization correction value DIGOP#i is calculated using the equations (26), (28) to (30), and the update of the integral value sum_σ_igop#i of the switching function is stopped.

sum_σ_igop#i(k)＝sum_σ_igop#i(k-1)                                                                                                            ... (30)

Above, when Lim_igop(Usl_igop#i(k))≠Usl_igop#i(k), or when DIGOP#i≥DIGPIAV#i, the update of the integral value sum_σ_igop#i of the switching function is suspended for the following reasons . That is, when Usl_igop#i(k) is a value outside the above-mentioned predetermined range and is limited to DIGOP_L or DIGOP_H, or when DIGOP#i≧DIGPIAV#i, the correction value DIGCMP#i described later is not set to In the case of DIGOP#i, if the update of the integral value su_σ_igop#i of the switching function is continued, the value thereof increases. As a result, when Usl_igop#i(k) returns to the above-mentioned specified range, or when DIGOP#i<DIGPIAV#i, so that the correction value DIGCMP#i is set to DIGOP#i, the enlarged switching function It takes a long time until the integral value sum_σ_igop#i returns to an appropriate value, during which time the control accuracy decreases, so it is necessary to avoid this.

In addition, in the present embodiment, the optimization controller 36 corresponds to the first control value calculation means, and the optimization correction value DIGOP#i corresponds to the first control value.

Then, in the correction value calculation unit 37, the correction value DIGCMP# is calculated using the following equation (31) or (32) based on the comparison result between the smoothing correction value DIGPIAV#i and the optimization correction value DIGOP#i calculated as described above. i.

·When DIGOP#i(k)≥DIGPIAV#i(k) is satisfied,

Digcmp#i (k) = digpiav#i (k) ..... (31)

·When satisfying DIGOP#i(k)<DIGPIAV#i(k),

Digcmp#i (k) = digop#i (k) ..... (32)

Referring to the above equations (31) and (32), it can be seen that the correction value DIGCMP#i is calculated as a value obtained by subjecting the smoothing correction value DIGPIAV#i to the limit processing with the optimization correction value DIGOP#i as the upper limit. . The reason for this will be described later. In addition, in the present embodiment, the correction value calculation unit 37 corresponds to the ignition timing determination means and the final control value calculation means, and the correction value DIGCMP#i corresponds to the final control value.

On the other hand, the basic ignition timing calculation unit 38 calculates the basic ignition timing IGBASE by searching the map shown in FIG. 6 based on the engine speed NE and the accelerator opening AP. In this map, the basic ignition timing IGBASE is set to a value at which the degree of advance angle decreases as the accelerator opening AP increases. This is because, the larger the accelerator opening AP, the more the engine 3 is in the high load region, and knocking (knocking) tends to occur, so it needs to be avoided.

In addition, the basic ignition timing IGBASE is set to a value at which the higher the engine speed NE is, the greater the degree of advance angle is in the low rotation region, and is set to a value at which the higher the engine speed NE is, the greater the degree of advance angle is in the high rotation region. The smaller the value, this is because knocking is not likely to occur in the low-rotation region, so the ignition timing is set to a value where the higher the engine speed NE, the greater the degree of advance angle, thereby increasing the combustion gas temperature and improving combustion efficiency. On the other hand, since knocking tends to occur in the high rotation range, the ignition timing is set to a value such that the higher the engine speed NE, the smaller the advance angle, thereby avoiding the occurrence of knocking. In addition, in this map, the base ignition timing IGBASE is set to a value on the retard side of MBT.

Then, in the adder 39, the ignition timing IGLOG#i is calculated using the following equation (33) from the basic ignition timing IGBASE calculated as described above, the identification signal value DIGID, and the correction value DIGCMP#i.

IGLOG#i(k)＝IGBASE(k)+DIGID(k)+DIGCMP#i(k) .....(33)

In addition, in the present embodiment, the basic ignition timing calculation unit 38 and the adder 39 correspond to ignition timing determination means and ignition timing calculation means.

Next, the ignition timing control process executed by the ECU 2 will be described with reference to FIG. 7 . This processing is processing for calculating the ignition timing IGLOG, and is executed in synchronization with the generation of the TDC signal.

In this process, first, in step 1 (abbreviated as "S1" in the figure; the same applies hereinafter), the count value Ccyl is incremented by 1. Then, go to step 2 and judge whether the count value Ccyl is greater than the value 4 or not.

When the judgment ends with "No", proceed to step 4 described later. On the other hand, when the judgment result of step 2 is "Yes", that is, when Ccyl>4, go to step 3, set the count value Ccyl to a value of 1, and then go to step 4. That is, after the count value Ccyl is incremented from the value 1 to the value 4, it is incremented from the value 1 again.

In step 4, the cylinder number value #i is set by searching the table shown in FIG. 8 based on the counter value Ccyl. Referring to this figure, it can be seen that when the count value Ccyl is set in the order of "1 → 2 → 3 → 4 → 1..." described above, the cylinder number value #i is set in the order of "#1 → #3 → # 4→#2→#1..." to set in order.

Then, in step 5, it is judged whether the lift failure flag F_LIFTNG is "1". The lift failure flag F_LIFTNG is set to "1" when a failure occurs in the variable valve lift mechanism 10 and is set to "0" when the variable valve lift mechanism 10 is normal in a determination process not shown.

When the judgment result is "No" and the variable valve lift mechanism 10 is normal, proceed to step 6 to judge whether the engine start flag F_ENGSTART is "1". In the judgment process not shown in the figure, the engine start flag F_ENGSTART is set by judging whether it is in the engine start control, that is, cranking, based on the engine speed NE and the ON/OFF signal of the IG/SW 25, Specifically, "1" is set when the engine start control is in progress, and "0" is set otherwise.

When the judgment result is "Yes" and the engine start control is in progress, go to step 7, set the ignition timing IGLOG#i to a predetermined start value Ig_crk (for example, BTDC10°), and then end this process.

On the other hand, if the result of the judgment in step 6 is "No" and the engine start control is not in progress, the process proceeds to step 8, where it is judged whether the accelerator opening AP is smaller than the predetermined value APREF. When the judgment result is "Yes" and the accelerator pedal is not depressed, the process proceeds to step 9, where it is judged whether the timer value Tast of the start-up timer is smaller than the predetermined value Tastlmt. The post-start timer is a timer for counting the elapsed time after the engine start control is completed, and is composed of an increment timer.

When the judgment result is "Yes" and Tast<Tastlmt, it is considered that the catalyst warm-up control should be executed, and the process proceeds to step 10 to calculate the catalyst warm-up value Ig_ast. Specifically, the catalyst warm-up value Ig_ast is calculated using a sliding mode control algorithm represented by the following equations (34) to (36).

$\mathrm{<span\; class="notranslate">\; Ig</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; ast</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; Ig</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; ast</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; base</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; Krch</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; ast</span>}<span\; class="notranslate">\; \&Center\; Dot;</span>\mathrm{<span\; class="notranslate">\; \&sigma;</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; ast</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span>$

$<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; Kadp</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; ast</span>}<span\; class="notranslate">\; \&Center\; Dot;</span>\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; k</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\mathrm{<span\; class="notranslate">\; \&sigma;</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; ast</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 34</span>}<span\; class="notranslate">\; )</span>$

σ_ast(k)＝Enast(k)+pole_ast·Enast(k-1)                                                                                                                                                                                                                                 ... (35)

Enast(k)＝NE(k)-NE_ast

In the above formula (34), Ig_ast_base represents a predetermined base ignition timing for catalyst warm-up (for example, BTDC 5°), and Krch_ast and Kadp_ast represent predetermined feedback gains. Also, σ_ast is a switching function defined by Equation (35). In this equation (35), pole_ast is a switching function setting parameter set to satisfy the relationship of -1<poie_ast<0, and Enast is a following error calculated according to equation (36). In Equation (36), NE_ast is a predetermined target rotation speed (for example, 1800 rpm) for warming up the catalyst. Using the above-described control algorithm, the catalyst warm-up value Ig_ast is calculated as a value at which the engine speed NE converges to the catalyst warm-up target speed NE_ast.

Then, the process proceeds to step 11, and after the ignition timing IGLOG#i is set to the catalyst warm-up value Ig_ast, this process ends.

On the other hand, when the judgment result of step 8 or 9 is "No", that is, when the accelerator pedal is depressed, or when Tast≧Tastlmt, go to step 12 and execute Pmi optimization and smoothing control processing.

Specifically, this Pmi optimization and smoothing control processing is executed as shown in FIG. 9 . First, in step 20, the basic ignition timing IGBASE is calculated. That is, as described above, the base ignition timing IGBASE is calculated by searching the map in FIG. 6 based on the engine speed NE and the accelerator opening AP.

Then, the process proceeds to step 21 to increment the count value Cdigid by 1, and then in step 22 it is judged whether the count value Cdigid is greater than a predetermined upper limit value Cdigid_max. When the judgment result is "No", the process proceeds to step 24 described later.

On the other hand, when the judgment result of step 22 is "Yes" and Cdigid>Cdigid_max, go to step 23, set the count value Cdigid to a value of 0, and then go to step 24. That is, after the count value Cdigid is incremented from the value 0 to the value Cdigid_max, it is reset to the value 0 and incremented again.

In step 24, as described above, the identification signal value DIGID is calculated by searching the table in FIG. 5 based on the counter value Cdigid.

Then, the process proceeds to step 25, and the estimated value Pmi_act#i indicating the mean effective pressure is calculated using the above formula (1). Thereafter, in step 26, the coefficient vector θ#i is calculated using the above expressions (14) to (25).

Then, in step 27, the MBT deviation EIGOP#i is calculated using the above formula (12), and then in step 28, the above formulas (26) to (29), or the formulas (26), (28) to (30 ) to calculate the optimal correction value DIGOP#i.

Then, in step 29, the smoothing correction value DIGPIAV#i is calculated using the above expressions (5) to (10), or using the expressions (5), (7) to (11). Then, in step 30, the correction value DIGCMP#i is calculated using the above formula (31) or (32).

Then, in step 31, the sum of the base ignition timing IGBASE calculated as described above, the identification signal value DIGID, and the correction value DIGCMP#i is set as the ignition timing IGLOG#i, and this process ends.

Returning to FIG. 7 , after the Pmi optimization and smoothing control processing in step 12 is executed as described above, this processing ends.

On the other hand, when the judgment result in step 5 is "Yes" and the variable valve lift mechanism 10 fails, the process proceeds to step 13 to calculate the value Ig_fs for failure. Specifically, the failure value Ig_fs is calculated using the sliding mode control algorithm of the following expressions (37) to (39).

$\mathrm{<span\; class="notranslate">\; Ig</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; fs</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; Ig</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; fs</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; base</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; Krch</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; fs</span>}<span\; class="notranslate">\; \&Center\; Dot;</span>\mathrm{<span\; class="notranslate">\; \&sigma;</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; fs</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; k</span>}<span\; class="notranslate">\; )</span>$

$<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; Kadp</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; fs</span>}<span\; class="notranslate">\; \&CenterDot;</span>\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; k</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\mathrm{<span\; class="notranslate">\; \&sigma;</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; fs</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 37</span>}<span\; class="notranslate">\; )</span>$

σ_fs(k)＝Enfs(k)+pole_fs Enfs(k-1)                                                                                                                                                                                                  ... (38)

Enfs(k)＝NE(k)-NE_fs                                                                                                ...(39)

In the above formula (37), Ig_fs_base represents a predetermined reference ignition timing for failure (for example, TDC±0°), and Krch_fs and Kadp_fs represent predetermined feedback gains. Also, σ_fs is a switching function defined by Equation (38). In this equation (38), pole_fs is a switching function setting parameter set to satisfy the relationship of -1<poie_fs<0, and Enfs is a following error calculated according to equation (39). In this formula (39), NE_fs is a predetermined target rotation speed at the time of failure (for example, 2000 rpm). Using the above-mentioned control algorithm, the value for failure Ig_fs is calculated as a value that causes the engine speed NE to converge to the above-mentioned failure-time target speed NE_fs.

Then, the process proceeds to step 14, and the ignition timing IGLOG#i is set to the value Ig_fs for failure, and this process ends.

Next, the intake air amount control process executed by the ECU 2 will be described with reference to FIG. 10 . This process is a process of calculating a lift control input Uliftin for controlling the intake air amount by the variable valve lift mechanism 10, and is executed at a predetermined control cycle ΔT (for example, 10 msec).

In this process, first, in step 40, it is judged whether or not the aforementioned lift failure flag F_LIFTNG is "1". When the judgment result is "No" and the variable valve lift mechanism 10 is normal, the process proceeds to step 41, and it is judged whether the above-mentioned engine start flag F_ENGSTART is "1".

If the judgment result is "Yes" and the engine start control is in progress, the process proceeds to step 42, where the start value Gcyl_cmd_crk of the target intake air amount is calculated by searching the table shown in FIG. 11 based on the engine water temperature TW.

In this table, in the range where the engine water temperature TW is higher than the predetermined value TWREF1, the lower the engine water temperature TW, the value Gcyl_cmd_crk for starting is set to a larger value, and in the range of TW≤TWREF1, the value for starting The value Gcyl_cmd_crk is set to a predetermined value Gcylref. This is because, when the engine water temperature TW is low, the frictional force of the variable valve lift mechanism 10 increases, and this is compensated for.

Then, in step 43, the target intake air amount Gcyl_cmd is set to the above-mentioned starting value Gcyl_cmd_crk. Thereafter, the process proceeds to step 44, and the target valve lift Liftin_cmd is calculated by searching the map shown in FIG. 12 based on the target intake air amount Gcyl_cmd and the engine speed NE. In the figure, Gcyl_cmd1 to 3 are predetermined values of the target intake air amount Gcyl_cmd satisfying the relationship of Gcyl_cmd1<Gcyl_cmd2<Gcyl_cmd3.

As shown in the figure, in this map, the larger the target intake air amount Gcyl_cmd is, and the higher the engine speed NE is, the larger the value of the target valve lift Liftin_cmd is set. This is because the larger the target intake air amount Gcyl_cmd is, and the higher the engine speed NE is, the greater the output required to the engine 3 is, and thus a larger intake air amount Gcyl is required.

Then, go to step 45 to calculate the lift control input Uliftin. Using the target value filtering type 2-degree-of-freedom response designation control algorithm shown in the following equations (40) to (43), the lift control input Uliftin is calculated as a value that makes the valve lift Liftin follow the target valve lift Liftin_cmd . In addition, in the following equations (40) to (43), each discrete data with a symbol (m) represents data that is sampled or calculated in synchronization with the above-mentioned control period ΔT.

$\mathrm{<span\; class="notranslate">\; Uliftin</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; Krch</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; lf</span>}<span\; class="notranslate">\; \&Center\; Dot;</span>\mathrm{<span\; class="notranslate">\; \&sigma;</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; lf</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; Kadp</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; lf</span>}<span\; class="notranslate">\; \&CenterDot;</span>\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 0</span>}}{\overset{\mathrm{<span\; class="notranslate">\; m</span>}}{\mathrm{<span\; class="notranslate">\; \&Sigma;</span>}}}\mathrm{<span\; class="notranslate">\; \&sigma;</span>}<span\; class="notranslate">\; \_</span>\mathrm{<span\; class="notranslate">\; lf</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; .</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; 40</span>}<span\; class="notranslate">\; )</span>$

σ_if(m)＝E_lf(m)+pole_lf·E_lf(m-1)                                                                                                                                                                                                                                           ... (41)

E_lf(m)＝Liftin(m)-Liftin_cmd_f(m) .....(42)

Liftin_cmd_f(m)=-pole_f_if Liftin_cmd_f(m-1)

+(1+pole_f_lf)Liftin_cmd(m) .....(43)

In the above formula (40), Krch_lf represents the predetermined arrival law gain, Kadp_lf represents the predetermined adaptive law gain, and σ_lf is the switching function defined by the formula (41). In this equation (41), E_lf is the deviation calculated by the equation (42), and pole_lf is a switching function setting parameter, which is set to a value within the range of -1<pole_lf<0. In addition, in Equation (42), Liftin_cmd_f is a filter value of the target valve lift, and is calculated using a target value filtering algorithm (primary delay filtering algorithm) shown in Equation (43). In the formula (43), pole_f_lf is a target value response designation parameter, and is set to a value within the range of -1<pole_f_lf<0.

As described above, in step 45, the lift control input Uliftin is calculated, and this process ends.

On the other hand, if the result of the judgment in step 41 is "No" and the engine start control is not in progress, the process proceeds to step 46, where it is judged whether the accelerator opening AP is smaller than the predetermined value APREF. When the judgment result is "Yes" and the accelerator pedal is not depressed, the process proceeds to step 47, where it is judged whether the counted value Tast of the timer after start is smaller than the predetermined value Tastlmt.

When the judgment result is "yes" and Tast<Tastlmt, it is considered that the catalyst warm-up control should be executed, and the process goes to step 48. According to the timing value Tast of the timer after starting and the engine water temperature TW, the map shown in Fig. 13 is retrieved. , calculate the catalyst warm-up value Gcyl_cmd_ast for the target intake air amount. In the figure, TW1 to TW3 represent predetermined values of the engine water temperature TW satisfying the relationship of TW1<TW2<TW3.

In this map, the lower the engine water temperature TW is, the larger the catalyst warm-up value Gcyl_cmd_ast is set. This is because the lower the engine water temperature TW is, the longer the time required for catalyst activation is, and thus the time required for catalyst activation is shortened by increasing the exhaust gas volume. In addition, in this map, while the count value Tast of the timer after start is small, the value Tast for the catalyst warm-up is set to a larger value as the count value Tast increases, and when the count value Tast increases to After a certain period, the larger the timer value Tast is, the smaller the catalyst warm-up value Gcyl_cmd_ast is set. This is because, as the catalyst warm-up control execution time elapses, the engine 3 is warmed up to lower the frictional force. In this case, if the amount of intake air is not reduced, the engine speed NE In a state where excessive retardation control is performed on the ignition timing while maintaining the target value, the combustion state becomes unstable, and it is necessary to avoid this.

Then, the process proceeds to step 49, where the target intake air amount Gcyl_cmd is set to the catalyst warm-up value Gcyl_cmd_ast. Afterwards, as described above, steps 44 and 45 are executed, and then this process ends.

On the other hand, when the judgment result of step 46 or 47 is "No", that is, when the accelerator pedal is stepped on, or when Tast≥Tastlmt, proceed to step 50, according to the engine speed NE and the accelerator opening AP, The normal value Gcyl_cmd_drv of the target intake air amount is calculated by searching the map shown in FIG. 14 . In the figure, AP1 to AP3 represent predetermined values of the accelerator opening AP satisfying the relationship of AP1<AP2<AP3.

In this map, the higher the engine speed NE or the larger the accelerator opening AP, the larger the normal value Gcyl_cmd_drv is set. This is because the higher the engine speed NE or the larger the accelerator opening AP, the larger the output required from the engine 3, and thus a larger amount of intake air is required.

Then, the process proceeds to step 51, and the target intake air amount Gcyl_cmd is set to the above-mentioned value Gcyl_cmd_drv for normal use. Thereafter, as described above, steps 44 and 45 are executed, and this process ends.

On the other hand, when the judgment result of step 40 is "Yes", and the variable valve lift mechanism 10 fails, proceed to step 52, set the lift control input Uliftin to the specified failure value Uliftin_fs, and then end This processing.

Next, simulation results (hereinafter referred to as "control results") of ignition timing control by the ignition timing control device 1 of the present embodiment will be described. Fig. 15 shows the control results under the following conditions: the basic ignition timing IGBASE of the four cylinders #1 to #4 is kept constant, the weighting coefficients w1 and w2 are set to w1=0.2, w2=0.8, from the start of the control to During the end period, the statistical processing values Pmi_ls#1 to #4 of the four cylinders satisfy Pmi_ls#1>Pmi_ls#2>Pmi_ls#3>Pmi_ls#4, and the statistical processing value Pmi_ls#3 of the third cylinder is set as the first The reference value Pmi_ls_min1 and the statistically processed value Pmi_ls#2 of the second cylinder are set as the second reference value Pmi_ls_min2, that is, so that Pmi_ls#3≥Pmi_1s#1−Eps_Pmi and Pmi_ls#2≥Pmi_ls#1−Eps_Pmi.

In the figure, DIG1 to 3 represent predetermined values of the correction value DIGCMP#i satisfying DIG1<DIG2<DIG3, Pmi_max#3 represents the maximum value of the indicated mean effective pressure of the third cylinder #3 under the estimated MBT condition, That is, it represents the maximum value of the indicated mean effective pressure attainable in the third cylinder #3.

Referring to this figure, it can be seen that first, in the fourth cylinder #4, the smoothing target value Piav_cmd is calculated by the weighted average calculation of the formula (4) as described above, and is compared with the statistically processed value Pmi_ls# of the fourth cylinder #4. Therefore, the smoothing correction value DIGPIAV#4 is calculated as a value that corrects the basic ignition timing IGBASE to be closer to the advance angle side than the estimated MBT. It can also be seen that since Pmi_ls#4 cannot reach Piav_cmd, DIGPIAV#4 is maintained at the upper limit value DIGPIAV_H after a sharp increase.

In this case, from the relationship between the operation speed of DIGPIAV #4 of the smoothing controller 33 and the operation speed of DIGOP #4 of the optimization controller 36, during the period from the control start time t0 to time t1, DIGOP #4≥DIGPIAV# 4 is established, so the correction value DIGCMP#4 is set as the smoothing correction value DIGPIAV#4, and after time t2, DIGOP#4<DIGPIAV#4 holds, so the correction value DIGCMP#4 is set as the optimal correction value DIGOP# 4. As a result, in the 4th cylinder #4, the ignition timing IGLOG#4 is controlled to the estimated MBT, and Pmi_ls#4 is controlled to converge to the maximum value of the indicated mean effective pressure of the 4th cylinder #4 under the estimated MBT condition.

Then, in the third cylinder #3, the smoothing target value Piav_cmd is calculated by the weighted average calculation of the formula (4), and the smoothing target value Piav_cmd is set to be higher than the statistically processed value Pmi_ls# of the third cylinder #3. 3, DIGOP#3≥DIGPIAV#3 holds true during the period from the control start time t0 to time t2, and the correction value DIGCMP#3 is set as the smoothing correction value DIGPIAV#3. Then, after time t2, DIGOP#3<DIGPIAV#3 holds true, and correction value DIGCMP#3 is set as optimum correction value DIGOP#3. As a result, in the 3rd cylinder #3, ignition timing IGLOG#3 is controlled to estimate MBT, and Pmi_ls#3 is controlled to converge to Pmi_max#3.

On the other hand, in the second cylinder #2, the smoothing target value Piav_cmd is calculated by the weighted average calculation of the expression (4), and the smoothing target value Piav_cmd is set to be higher than the statistically processed value Pmi_ls of the second cylinder #2. #2 small value. Therefore, DIGOP#2≧DIGPIAV#2 always holds, and thus the correction value DIGCMP#2 is set as the smoothing correction value DIGPIAV#2. As a result, in the second cylinder #2, the ignition timing IGLOG#2 is controlled to a value on the retard side of the estimated MBT, and Pmi_ls#2 is controlled to converge to Pmi_max#3.

Similarly, in the first cylinder #1, DIGOP#1≧DIGPIAV#1 is always established for the reason described above, so the correction value DIGCMP#1 is set as the smoothing correction value DIGPIAV#1. As a result, also in the first cylinder #1, the ignition timing IGLOG#1 is controlled to a value on the retard side of the estimated MBT, and Pmi_ls#1 is controlled to converge to Pmi_max#3.

As described above, according to the ignition timing control device 1 of the present embodiment, the smoothing correction value DIGPIAV#i is used as the statistically processed values Pmi_ls#1 to #4 for controlling the ignition timing IGLOG#i so that the indicated mean effective pressures of all cylinders follow One correction value of the smoothing target value Piav_cmd is calculated, and by using this smoothing correction value DIGPIAV#i to correct the ignition timing IGLOG#i, the ignition timing IGLOG#i can be controlled so that there is no variation in the combustion state between cylinders , that is, there is no thermal deviation. In this way, it is possible to avoid rotation speed fluctuations, vibrations, and the like due to variations in the combustion state among the cylinders. In particular, when the ignition timing control device 1 is applied to an internal combustion engine 3 prone to variation in the combustion state by providing the variable valve lift mechanism 10 and the like, the above advantages can be more effectively obtained.

In addition, since the smoothing target value Piav_cmd is calculated as the weighted average of the first and second reference values Pmi_ls_min1 and Pmi_ls_min2, which are statistically processed values of two cylinders, it is possible to avoid setting the smoothing target value Piav_cmd to be consistent with all cylinders. The statistically processed values Pmi_ls#1 to #4 deviate greatly from each other, so that the followability of the statistically processed value Pmi_ls#i to the smoothing target value Piav_cmd can be improved, and the occurrence of rotational speed fluctuations can be avoided. Through the above, the running performance can be improved.

Moreover, since the first and second reference values Pmi_ls_min1 and Pmi_ls_min2 used in the calculation of the smoothing target value Piav_cmd are set to the minimum value and the second small value, so even when there is a cylinder such that Pmi_ls#i<Pmi_ls_max-Eps_Pmi holds true, that is, when there is a cylinder whose combustion state is extremely worse than other cylinders, the combustion state can be better than that of the cylinder The ignition timing IGLOG#i is controlled based on the statistically processed values Pmi_ls#i of the two cylinders so that there is no combustion state deviation. As a result, unlike the case where the smoothing target value Piav_cmd is calculated based on the cylinder whose combustion state is extremely worse than other cylinders, it is possible to avoid the high temperature of the exhaust gas and the occurrence of misfires, etc., and to avoid damage to the exhaust system and the catalyst. .

Furthermore, since the statistically processed value Pmi_ls#i is calculated by performing predetermined ordinal statistical processing [Equations (2), (3)] on the estimated value Pmi_act#i calculated from the cylinder internal pressure Pcyl#i, it is different from directly using Compared with the case of the estimated value Pmi_act#i, the statistically processed value Pmi_ls#i can be regarded as a value that is not easily affected by fluctuations in the cylinder internal pressure Pcyl#i, that is, a value that suppresses the influence of noise due to fluctuations. It is calculated that the ignition timing can be controlled using such a statistically processed value Pmi_ls#i. As a result, by suppressing the influence of noise caused by fluctuations in the cylinder internal pressure Pcyl#i, the control accuracy of the ignition timing IGLOG#i can be improved accordingly.

Furthermore, the optimization correction value DIGOP#i is calculated as a correction value for controlling the ignition timing IGLOG#i of each cylinder #i to the estimated MBT, and the correction value DIGOP# is optimized by performing smoothing correction value DIGPIAV#i The correction value DIGCMP#i is calculated by limit processing where i is a limit value on the advance angle side, and the base ignition timing IGBASE is corrected using the optimization correction value DIGOP#i to calculate the ignition timing IGLOG#i. Therefore, in the ignition timing control of each cylinder, the ignition timing control for controlling the ignition timing IGLOG#i of each cylinder to the estimated MBT and the method for making the statistical processing value Pmi_ls#i follow the smoothing target are executed simultaneously. The situation of the ignition timing control of the value Piav_cmd is different, the interference of the two ignition timing controls can be avoided, and the switching of the correction value DIGCMP#i from the smoothing correction value DIGPIAV#i to the optimized correction value DIGOP#i can be smoothly performed . In this way, the occurrence of torque steps can be avoided, and the running performance can be improved.

Furthermore, since the correction value DIGCMP#i is calculated by subjecting the smoothing correction value DIGPIAV#i to the limit processing with the optimal correction value DIGOP#i as the limit value on the advance angle side, in any cylinder, even the smoothing correction value When DIGPIAV#i is set to a value closer to the advance angle side than the optimal correction value DIGOP#i, by limiting the correction value DIGCMP#i of the cylinder to the optimal correction value DIGOP#i, the optimal correction value can also be adjusted according to the optimal correction value. DIGOP#i calculates the ignition timing IGLOG#i, so that the ignition timing IGLOG#i of the cylinder is controlled to be the estimated MBT, and the statistically processed value Pmi_ls#i indicating the mean effective pressure is controlled to be the maximum value. That is, the highest combustion efficiency can be ensured. In this way, the running performance can be further improved.

In addition, since the smoothing target value Piav_cmd is calculated as the weighted average of the first and second reference values Pmi_ls_min1 and Pmi_ls_min2 as described above, and the correction value DIGCMP#i is determined as described above, the statistically processed value Pmi_ls In the cylinder where #i is set as the first reference value Pmi_ls_min1, the correction value DIGCMP#i is set as the optimization correction value DIGOP#i, and the ignition timing IGLOG#i is controlled so that the probability of estimated MBT increases. As a result, in addition to the first reference value Pmi_ls_min1 cylinder, in the cylinder where Pmi_ls#i<Pmi_ls_max-Eps_Pmi is satisfied, the ignition timing IGLOG#I is also controlled to estimate MBT, so the combustion efficiency of the engine 3 as a whole can be improved more effectively. .

Furthermore, since the smoothing correction value DIGPIAV#i is a value for controlling the ignition timing IGLOG#i of each cylinder so that all four statistically processed values Pmi_ls#1 to #4 follow the smoothing target value Piav_cmd, the engine 3 When the cylinder with the first reference value Pmi_ls_min1 is switched, the smoothing target value Piav_cmd suddenly changes, and the smoothing correction value DIGPIAV#i may also change suddenly. In contrast, according to the ignition timing control device 1, since the two-degree-of-freedom The control algorithm [Equation (5) to (10)] of the control algorithm [Equation (5) to (9)] calculates the smoothing correction value DIGPIAV#i, so by properly setting the filter characteristics of the target value filter algorithm, even in When the smoothing target value Piav_cmd changes suddenly, the smoothing correction value DIGPIAV#i can also be avoided from changing suddenly. As a result, it is possible to avoid the occurrence of a torque step caused by a sudden change in the smoothing correction value DIGPIAV#i, and to further improve the drivability.

In addition, the calculation method of the correction value DIGCMP#i is not limited to the method using the above-mentioned formulas (31) and (32), and the following method may be used. That is, among the cylinders whose statistical processing value Pmi_ls#i is set to the second reference value Pmi_ls_min2 and the cylinders whose statistical processing value Pmi_ls#i is larger than the second reference value Pmi_ls_min2, the correction value is calculated as shown in the following equation (44). DIGCMP#i is set as the smoothing correction value DIGPIAV#i.

DIGCMP#i(k)＝DIGPIAV#i(k)                                                                                                                                                                                                                                               ... (44)

On the other hand, among the cylinders other than the above, that is, the cylinders whose statistical processing value Pmi_ls#i is set to the first reference value Pmi_ls_min1 and the cylinders whose statistical processing value Pmi_ls#i is smaller than the first reference value Pmi_ls_min1, the following As shown in equation (45), the correction value DIGCMP#i is set as the optimum correction value DIGOP#i.

DIGCMP#i(k)＝DIGOP#i(k)                                                                                                                                                                                                                                                      ... (45)

Next, the simulation results (hereinafter referred to as "control results") of the ignition timing control in the case where the correction value DIGCMP#i is calculated using the above method will be described. Fig. 16 shows an example of the control result of the above-mentioned method. The basic ignition timing IGBASE, weighting coefficients w1, w2 and statistical processing values Pmi_ls #1-#4 of the four cylinders #1-#4 are set to be the same as those of the above-mentioned embodiment. The example is the same. In addition, FIG. 17 shows that in the ignition timing control method of the embodiment, by setting the target value response designation parameter R_piav of the smoothing controller 33 to an extremely small value close to the value 0, the ignition timing is intentionally extreme. An example of the control result when the following speed of the statistical processing value Pmi_ls#i to the smoothing target value Piav_cmd is greatly increased.

First, referring to the comparative example in FIG. 17, it can be seen that, in the fourth cylinder #4, as in the control result example of FIG. The maximum value of the indicated mean effective pressure at the time of estimating the MBT of the fourth cylinder #4, but the statistically processed values Pmi_ls#1 to #3 of the other cylinders converge to a value higher than that attainable in the third cylinder #3. Maximum value of effective pressure Pmi_max#3 low value. This is because, as described above, since the follow-up speed of the statistical processing value Pmi_ls#i to the smoothing target value Piav_cmd is extremely increased, the optimization correction value DIGOP#3 of the third cylinder #3 is lower than the smoothing correction value DIGPIAV# In the state of 3, the statistically processed values Pmi_ls#1 to #3 of the first to third cylinders are smoothed. Since this state occurs when the following speed of the statistical processing value Pmi_ls#i to the smoothing target value Piav_cmd is extremely increased in the smoothing controller 33, for example, when the weighting coefficients w1, w2 are set such that w1> This state also occurs at an inappropriate value such as >w2 holds.

In contrast, in the example of the control result shown in FIG. 16 , in the fourth cylinder #4, DIGOP #4≥DIGPIAV #4 holds true from the control start time t10 to time t11, but after time t11, DIGOP#4<DIGPIAV#4 holds true, but as shown in the above formula (45), since the correction value DIGCMP#4 is kept as the optimum correction value DIGOP#4, the ignition timing IGLOG #4 is controlled to estimate the MBT, and Pmi_ls#4 is controlled to converge to the maximum value of the indicated mean effective pressure of the 4th cylinder #4 at the time of estimating the MBT.

Also, in the third cylinder #3, since the correction value DIGCMP#3 is held as the optimum correction value DIGOP#3, the ignition timing IGLOG#3 is controlled to be the estimated MBT as in the example of the control result of FIG. 15 described above, and Pmi_ls#3 is controlled to converge to the maximum value Pmi_max#3 of the indicated mean effective pressure of the third cylinder #3 at the time of estimating the MBT.

On the other hand, it can be seen that in the first and second cylinders, the correction values DIGCMP #1 and #2 are held as the smoothing correction values DIGPIAV #1 and #2, respectively, and the statistically processed values Pmi_ls #1 and #2 converge at The maximum value Pmi_max#3 of the indicated mean effective pressure attainable by the third cylinder #3. That is, it can be seen that the combustion efficiency of the engine 3 as a whole is effectively improved.

As described above, even when the modified examples of the above formulas (44) and (45) are used, the same effects as those of the embodiment can be obtained. In addition, in the case of this modified example, due to a change in the operating state of the engine 3, the statistical processing value Pmi_ls#i is set as the first reference value Pmi_ls_min1 for the cylinder and the statistical processing value Pmi_ls#i is set as When the cylinder of the second reference value Pmi_ls_min2 is switched, the correction value DIGCMP#i is abruptly switched between the smoothing correction value DIGPIAV#i and the optimization correction value DIGOP#i. A little torque differential.

In addition, the embodiment uses the cylinder internal pressure sensor 22 as the cylinder internal pressure detection device, however, the cylinder internal pressure detection device is not limited thereto, as long as it can detect the cylinder internal pressure Pcyl#i.

Also, the embodiment uses the statistically processed value Pmi_ls#i indicating the mean effective pressure as the work amount parameter, but the work amount parameter is not limited thereto, as long as it is a value calculated from the in-cylinder pressure and represents work caused by combustion in the cylinder. The value of the quantity is enough. For example, as the work amount parameter, an indicated output, a heat generated, etc. may be used, or an estimated value Pmi_act#i indicating an average effective pressure may be directly used.

Furthermore, the embodiment uses the first and second reference values Pmi_ls_min1 and Pmi_ls_min2 as reference values, and calculates the smoothing target value Piav_cmd through their weighted average calculation. However, the calculation method of the smoothing target value Piav_cmd is not limited to this, as long as statistical processing is used One of the values Pmi_ls#i may be used as a reference value, and a smoothing target value Piav_cmd may be calculated from the reference value. For example, when the minimum value of the four statistically processed values Pmi_ls#i is smaller than Pmi_ls_max-Eps_Pmi, this minimum value may be set as a reference value, and the smoothing target value Piav_cmd may be calculated from this minimum value. In addition, the smoothing target value Piav_cmd may be calculated by weighted average calculation of such a minimum value and the second smallest value.

In addition, the embodiment uses the sliding mode control algorithm [Formulas (5) to (9)] as the prescribed follow-up control algorithm, but the prescribed follow-up control algorithm is not limited thereto, as long as it is used to control so that the statistically processed value Pmi_ls#i follows An algorithm for smoothing the filter value Piav_cma_f of the target value is sufficient. For example, as the following control algorithm, other response-specific control algorithms such as back-stepping control algorithm can be used, or PID control algorithm (Proportional, Integral and Derivative Control Algorithm, proportional, integral, differential control algorithm can be used) ) and other conventional feedback control algorithms.

Furthermore, the embodiment applies the ignition timing control device of the present invention to a 4-cylinder internal combustion engine, but the ignition timing control device of the present invention is not limited thereto, and may be applied to a multi-cylinder internal combustion engine with 2 or more cylinders.

Furthermore, in the embodiment, the ignition timing control device of the present invention is applied to a vehicle internal combustion engine, but the ignition timing control device of the present invention is not limited thereto, and may be applied to a general-purpose internal combustion engine. For example, the ignition timing control device of the present invention can be applied to internal combustion engines for ships, power generation, and the like.

The above is the description of preferred embodiments of the present invention, and it should be understood by those skilled in the art that various changes can be made without departing from the spirit and scope of the present invention.

Claims (8)

1. the igniting correct timing controller of an internal-combustion engine is characterized in that, has:

Press detection device in the cylinder, it detects pressure in a plurality of cylinders respectively with as pressing in a plurality of cylinders;

Acting amount parameter calculation apparatus, it calculates a plurality of acting amount parameters of expression by the caused acting amount of burning in described a plurality of cylinders respectively according to pressing in these detected a plurality of cylinders;

The desired value computing device, it is 1 reference value that is set at these a plurality of acting amount parameters in these a plurality of acting amount parameters of calculating, and according to this reference value, calculates 1 desired value of the target that becomes described a plurality of acting amount parameters; And

The ignition timing determination device, it determines the ignition timing of described a plurality of cylinders to make described 1 desired value of described a plurality of acting amount parameter tracking respectively.

2. the igniting correct timing controller of internal-combustion engine according to claim 1 is characterized in that, described reference value is set to the minimum value in described a plurality of acting amount parameter.

3. the igniting correct timing controller of internal-combustion engine according to claim 1, it is characterized in that described reference value is set to and is being the upper limit with the maximum value in described a plurality of acting amount parameters, is being the minimum value in the acting amount parameter in the scope of lower limit with the value than the little specified value of this maximum value.

4. the igniting correct timing controller of internal-combustion engine according to claim 1, it is characterized in that, described acting amount parameter calculation apparatus is according to pressing the value of being untreated of calculating described a plurality of acting amount parameters respectively in described a plurality of cylinders, and, calculate described a plurality of acting amount parameter by this value of being untreated being implemented the statistical process of type successively of regulation.

5. the igniting correct timing controller of internal-combustion engine according to claim 1 is characterized in that, described ignition timing determination device has:

The 1st controlling value computing device, it calculates the described ignition timing that is used at described each a plurality of these each cylinder of cylinder control, makes the acting amount parameter of this each cylinder be the 1st maximum controlling value;

The 2nd controlling value computing device, it calculates the described ignition timing that is used at described this each cylinder of each cylinder control, makes the 2nd controlling value of described 1 desired value of acting amount parameter tracking of this each cylinder;

Final controlling value computing device, its limit by the limiting value that to implement with described the 1st controlling value to described the 2nd controlling value be advance side is handled, and calculates final controlling value; And

The ignition timing computing device, it calculates the described ignition timing of described each cylinder according to this final controlling value of calculating.

6. the igniting correct timing controller of internal-combustion engine according to claim 5 is characterized in that, the described reference value of described desired value computing device by described a plurality of acting amount parameters and the weighted mean of the value bigger than this reference value are calculated described 1 desired value.

7. the igniting correct timing controller of internal-combustion engine according to claim 3 is characterized in that, described ignition timing determination device has:

The 1st controlling value computing device, it calculates the described ignition timing that is used at described each a plurality of these each cylinder of cylinder control, makes the acting amount parameter of this each cylinder be the 1st maximum controlling value;

The 2nd controlling value computing device, it calculates the described ignition timing that is used at described this each cylinder of each cylinder control, makes the 2nd controlling value of described 1 desired value of acting amount parameter tracking of this each cylinder; And

The ignition timing computing device, it calculates the described ignition timing of described acting amount parameter smaller or equal to the cylinder of described reference value according to described the 1st controlling value, calculates the described ignition timing of described acting amount parameter greater than the cylinder of described reference value according to described the 2nd controlling value.

8. according to the igniting correct timing controller of claim 5 or 7 described internal-combustion engines, it is characterized in that, described the 2nd controlling value computing device use comprise will regulation the desired value filtering algorithm and the control algorithm of following the 2DOF control algorithm that control algorithm makes up of regulation, calculate described the 2nd controlling value.

Images